18.1 Vortical Structures in Convective Boundary Layers and Implications for the Initiation of Deep Convection
Total Page:16
File Type:pdf, Size:1020Kb
18.1 VORTICAL STRUCTURES IN CONVECTIVE BOUNDARY LAYERS AND IMPLICATIONS FOR THE INITIATION OF DEEP CONVECTION Katharine M. Kanak School of Meteorology, University of Oklahoma, Norman, Oklahoma 1. INTRODUCTION 2.2. Experiment Design The purpose of this study is to identify coherent structures in the convective boundary layer, in To explore the coherent structures for convective particular, vertical vortices of scales ranging from dust boundary layers, in particular vertical vortices, several devil-scale to misocyclones-scale and to compare numerical experiments have been carried out (Table simulation results to observations. I). For the vertical coordinate where more than one Observations show that dust devils have been value is given for the grid spacing a vertically reported to be favored in environments characterized stretched coordinate has been used. It has been by little or no ambient winds (about less than 5 m s-1; computationally impossible to date to model both the Webb 1963; Sinclair 1969). Morton (1966) states that dust devil-scale and the misocyclone scale wind speeds greater than 7-10 m s-1 will break up a simultaneously, thus the problem has been separated dust devil. In addition, too much wind would modify into two parts. The first three simulations in Table 1 (Moeng and Sullivan 1994) the cellular convective are larger-scale, coarser resolution simulations that pattern to which the vertical vortices appear to be are intended to examine the larger-scale vertical integrally tied (Willis and Deardorff 1979; Mason vortices, such as misocyclones, that arise in the 1989; Kanak et al. 2000, hereafter KLS2000; Kanak cellular convective patterns themselves. The higher 2005). The presence of mean winds or wind shears resolution simulations (6M, 2.75M and 2M) were would provide a more obvious source of vorticity. So designed to determine whether or not dust devil-scale a more intriguing question is in such environments vortices could be simulated with realistic mean winds are negligible, from where does the observational physical characteristics. Vortices did vorticity come? form which generally compared well with observations Results are presented from numerical simulations (e.g. Kanak 2005). of the convective boundary layer (CBL) without imposed ambient winds that range from dust devil- 2.75 scale O(10 m) up to a “parent cyclone”-scale, which in 40M 35M 30M 6M 2M M some cases is on the order of the misocyclone scale O(100) m. The relationship between these vortices, !x their structure and possible role in CBL processes will 40 35 30 6 2.75 2 be presented. =!Y 10.5- 3- 2.5- 2. 2.METHODOLOGY z 20 25 ! 80.3 22 24 1-24 2.1. Numerical Model Description Lx 5600 3000 4320 1004 1018 740 =Ly The Kanak’s System for Atmospheric Simulation (KANSAS) is a dry, three-dimensional, fully Lz 2000 2100 3600 1400 1200 1200 compressible, nonhydrostatic numerical model that integrates the Navier-Stokes equations. The model Table 1. Summary of numerical experiments equations are presented in Kanak (1999) and are cast on an Arakawa C-grid. The velocity and scalar 3. RESULTS AND DISCUSSION advection is represented by a second-order quadratic conserving “box” finite difference spatial scheme 3.1. Vertical Structure (Kurihara and Holloway 1967), The subgrid turbulence closure is a first-order Smagorinsky-Lilly Vertical vortices may be manifest as dust devils (Smagorinsky 1962) scheme. A fourth-order in the presence of a visible tracer, such as dust, numerical filter is employed. There are no prescribed although with the increase of remote sensing mean ambient winds and the surface of the domain is instrumentation, evidence of invisible vertical vortices heated with a constant flux condition. The lateral is increasing (e.g., MacPherson and Betts 1997; boundaries are periodic, and the upper and lower KLS2000; Bos et al. 2005; Markowski and Hannon boundaries are rigid. The lower boundary is semi- 2006). In some cases, smaller diameter dust devils slip. are observed to be embedded within larger-scale “parent” circulations that are evident in observations By the next time sampling (Fig. 1b), 6 min later, the (e.g., KLS2000, Markowski and Hannon 2006). This dust devil was gone. This Fig. suggests that in some more complex vertical structure may be important in cases, the parent cyclones to dust devils may be the vertical transports of particulates to the free same thing as misocyclones. If so, this would provide atmosphere. In addition, larger-scale vertical vortices, a structural link between dust devils and larger-scale such as the parent cyclones, can be of misocyclone vortices. scale and in that case may have influence on deep What is the nature of this link or relationship convective initiation (e.g., Wilson et al. 1992; between the small-scale dust devil vortex and the Kingsmill 1995; Lee et al. 2000; Pietrycha and larger-scale parent/misocyclone? Sinclair (1966) Rasmussen 2004; Markowski and Hannon 2006; presents the relationship using sailplane data Arnott et al. 2006). (reproduced here as Fig. 2). In Fig. 2, the dust devil An example of a parent-scale vortex can be column is the very shallow narrow light grey shaded found in IHOP data. Figure 1 shows a horizontal region in the lowest 100’s of meters with a diameter of plane of the vertical vorticity at 100 m height and the O(10) m. The disturbed airflow region, as denoted by vertical velocity at 1000 m height (reproduced from departures in temperature and vertical velocity, Markowski and Hannon 2006). The white dashed line extend well up into the boundary layer, even as high is a mesoscale convergence zone. Local vorticity as three or more km and as wide as four km diameter centers that range from 1-4 km in diameter exist and (misocyclone scale). This Figure is a composite of these vorticity centers are said to have duration of 1-2 data from 14 dust devils that have been corrected for hrs. Numerous dust devils were reported on this date vortex tilt and sailplane motion. This Figure and near this time, but one in particular was reported represents a key point of this paper; that vortices with to co-exist with one of the larger-scale vorticity varying diameters exist in the CBL and may be centers (inside the white box). responsible for important CBL processes such as vertical transports. a) b) Dust Devil Fig. 2. (From Sinclair 1966). Composite data from 14 sailplane penetrations of dust devil’s upper structure. Light grey shading denotes the height and width of a dust devil column near the surface. UW (DW) denotes upwind (downwind) of the dust devil based on translation direction. Disturbed wind field aloft has horizontal diameter up to 10 times that of the dust column. Next, does a similar pattern appear in numerical simulations of the CBL? Figure 3 shows two XY Fig. 1. XY Cross-sections of vertical vorticity panels of horizontal velocity vectors from the 40M (shaded) at 0.1 km and vertical velocity simulation at 4140s and at a) z = 10 m and b) z =590 (contoured) at 1.0 km. m. Here an anti-cyclonic vortex of about 100 m cyclonic vortex, particularly on the northeast side. diameter exists at the low level and aloft, a 500 m This is a common feature observed by Sinclair (1973) vortex (misocyclone scale) of the same sign exists. and KLS2000. The maximum negative vertical The vortex is apparently broadened with height vorticity amplitude is -1.4 s-1. (similar to Fig. 2). Aloft there is considerably more At a height of 98 m (Fig. 4b), the horizontal swirling motion, but because it is spread over larger velocity vectors show that the anti-cyclonic vortex has cross-sectional areas, the largest amplitudes of indeed broadened with height to a diameter of vertical vorticity are at the lowest model level. approximately 36 m. This is still considered a dust a) a) b) b) Fig. 3. 40M results, XY plane of horizontal Fig. 4. 2.75M results, XY planes of horizontal velocity vectors at 4140s. a) at z = 10 m. Diameter velocity vectors at 1700s. a) at z = 2.5 m. Diameter of anti-cyclonic vortex is approximately 100 m. b) of anti-cyclonic vortex is approximately 19 m. b) At z = 590 m. Diameter of broadened circulation is At z = 97.5 m. Diameter of broadened circulations approximately 500 m. is approximately 36 m. Next the higher resolution simulations are devil-scale vortex size. Perhaps a deeper domain examined to see whether or not vortices broaden with might be required to allow this vortex to broaden height in those simulations as well. Figure 4 shows further. The vorticity center has spread out laterally horizontal velocity vectors from the 2.75M simulation. -1 and the magnitude has decreased to –0.8 s , likely The diameter of the main anti-cyclonic vortex in Fig. owing to the available vertical vorticity being spread 4a is approximately 19 m and there is significant over a larger horizontal cross-section. cyclonic vorticity on periphery of the main anti- Thus, dust devil-scale vertical vortices may be associated with misocyclone scale vortices aloft. a) Large eddy simulation has shown that misocyclones can form in the absence of mesoscale convergence zones, but rather in the cellular convective pattern itself. In addition, observations and simulations show that dust devil-scale vortices may be integrally tied to convective circulations that broaden with height and may influence CBL processes. 3.2. Coherent Structures and Vertical Transports In 1988, Hunt et al. presented a schematic diagram (reproduced here as Fig. 5a,b) of the convective boundary layer. In particular, they showed a larger thermal updraft that extends over the depth of the boundary layer that included small shearing eddies and merging small plumes in the base of the thermal (Fig.